assessment of the contribution of agroforestry to poverty alleviation in lushoto district

24 Pages·2012·0.51 MB·English

July IJPSS Volume 2, Issue 7 ISSN: 2249-5894

2012

_________________________________________________________

Productivity of the Agroforestry Systems
and its Contribution to Household Income
among Farmers in Lushoto District, Tanzania

Baltazar M.L. Namwata*
Zacharia S. Masanyiwa*
Omari B. Mzirai*

__________________________________________________________
Abstract
This paper investigates the productivity of the agroforestry systems and its contribution to household
income among farmers in Lushoto District, Tanga, Tanzania. Specifically, the study aimed to determine
and compare the level of household’s farm production and net income between farmers practicing and not
practicing agroforestry. A total of 134 respondents from four villages in Soni and Ubiri wards were
involved. Data were collected using structured questionnaire, focus group discussion and through non-
participant observation. Distribution of responses, central tendency and dispersion, and multiple linear
regression analyses were carried using the Statistical Package for Social Science (SPSS) computer
software. Results indicate that farmers practicing agroforestry had significantly higher contribution to the
household’s level of farm production and net income than those who were not practicing agroforestry.
Given the average farm size of 3.1 ha, 2.3 cows and 9.2 chicken, the annual production for farmers
practicing agroforestry was 425.9 kg for maize, beans 225.7 kg, coffee 101.1kg, and 163.9 bunches of
banana, 999.12 litres and 373.5 eggs compared to 342.6 kg of maize, 202.1 kg of beans, 75 kg of coffee,
108 bunches of banana, 1120.6 litres of milk and 338.6 eggs for farmers not practicing agroforestry. The
average household annual net income was Tshs 664,992 and 547,608 for farmers practicing and not
practicing agroforestry respectively. The income per capita was Tshs 100,756 for farmers practicing and
Tshs 82,971 for farmers’ not practicing agroforestry. However, the level of household farm production
and net income was generally lower compared to most findings from other agroforestry systems. There

* Institute of Rural Development Planning, P.O. Box 138, Dodoma.
A Monthly Double-Blind Peer Reviewed Refereed Open Access International e-Journal - Included in the International Serial Directories
Indexed & Listed at: Ulrich's Periodicals Directory ©, U.S.A., Open J-Gage, India as well as in Cabell’s Directories of Publishing Opportunities, U.S.A.
International Journal of Physical and Social Sciences
http://www.ijmra.us
369
July IJPSS Volume 2, Issue 7 ISSN: 2249-5894

2012

_________________________________________________________

was partial adoption of the agroforestry technologies by some farmers (some households had few numbers
of trees/shrubs for timber, fodder, fruits, firewood etc) and poor management which resulted from
constraints like inadequate extension services, higher prices of most of the agricultural inputs, low soil
fertility and unattractive producer prices. Therefore, overcoming these constraints could improve and
probably sustain productivity of the agroforestry systems and its contribution to household income.
Key words: Agroforestry systems, crop production, household income, net income

Introduction
Kitalyi et al (2009) reported that Tanzania is listed among the thirteen African countries worst
affected by climate change impacts and vulnerability, and having the least adaptive capacities. A
review of the status of Tanzania’s Agricultural Sector Development Program notes that the
country is lagging in achieving its targets on reducing poverty and in achieving the Millennium
Development Goals due to low agricultural productivity. Low agricultural productivity is mainly
due to low and declining soil fertility. Soil fertility is low and declining due to reduced use of
inorganic fertilizers and continuous cropping. The resource poor farmers can not afford the
fertilizers in sufficient quantities and sometimes do not apply the fertilizer at all on some of the
lands. This is mainly because the farmers can not afford the fertilizers especially after the
introduction of the Economic Structural Adjustment Programme in 1989 to 1991 when fertilizer
subsidies were withdrawn. Tanzania faces the challenge of revitalizing her agricultural sector by
improving the natural resource base: soil, water and biodiversity. Agroforestry, the integration of
trees in agricultural landscapes, offers robust options to improve productivity and achieve
environmental sustainability.

Incorporating agroforestry systems into national agricultural development programmes offers
more affordable and sustainable sources of soil nutrients through deep soil extraction and nitrogen
fixation. On the other hand, practicing agroforestry or integrating agroforestry into the farming
systems can potentially improve the livelihoods of farmers through improved agricultural
productivity (SECAP, 1991; Semgalawe, 1998). Young (1989) defines agroforestry as a
collective name for land use systems and practices in which woody perennials (trees, shrubs etc)
A Monthly Double-Blind Peer Reviewed Refereed Open Access International e-Journal - Included in the International Serial Directories
Indexed & Listed at: Ulrich's Periodicals Directory ©, U.S.A., Open J-Gage, India as well as in Cabell’s Directories of Publishing Opportunities, U.S.A.
International Journal of Physical and Social Sciences
http://www.ijmra.us
370
July IJPSS Volume 2, Issue 7 ISSN: 2249-5894

2012

_________________________________________________________

are grown in association with herbaceous plants (crops, pastures) and/or livestock in a spatial
arrangement, rotation or both and in which there are both ecological and economic interactions
between the trees and non-tree components of the system. Tanzania is home to several traditional
agroforestry systems that have been in practice for hundreds of years. Some have been
documented: the Chagga home-gardens, the related Mara region home-gardens known as
Obohochere and the traditional Wasukuma silvopastoral system called Ngitili and those of
Kagera. One outstanding aspect of these traditional methods is the use of multi-layered systems
with a mixture of annual and perennial plants, which imitate natural ecosystems (O’kting’ati,
1985; Rugalema, 1992).

In effect, many projects with various packages of agroforestry technologies have been launched in
various parts of the country. Some of these projects include Soil Erosion Control and
Agroforestry Project (SECAP) in Lushoto, Soil Conservation and Agroforestry Project Arusha
(SCAPA), “Hifadhi Mazingira” (HIMA), “Hifadhi Ardhi Dodoma” (HADO), Land Management
Programme (LAMP), Handeni Integrated Agroforestry Project (HIAP), Hifadhi Ardhi Shinyanga
(HASHI) to mention a few. A lot have been done by these projects; starting from the point of
sensitization to integration of various agroforestry technologies into the poor rural farming
systems (Kerkhof, 1990; Johansson, 2001).

In Lushoto district where the SECAP project has been implemented since 1981, previous studies
have shown that, the project has been successful in improving the livelihood strategies of farming
households (Kerkhof, 1990, SECAP, 1999; Johansson, 2001). While the contribution of
agroforestry to poverty reduction is well documented in many places, there is paucity of
information on whether or not agroforestry systems significantly contribute to the farm
productivity and household income in Lushoto District. Therefore, this paper examined the
productivity of the agroforestry systems and its contribution to household income in the district.

A Monthly Double-Blind Peer Reviewed Refereed Open Access International e-Journal - Included in the International Serial Directories
Indexed & Listed at: Ulrich's Periodicals Directory ©, U.S.A., Open J-Gage, India as well as in Cabell’s Directories of Publishing Opportunities, U.S.A.
International Journal of Physical and Social Sciences
http://www.ijmra.us
371
July IJPSS Volume 2, Issue 7 ISSN: 2249-5894

2012

_________________________________________________________

Materials and Methods
This study was carried out four selected villages namely Ngulwi and Ubiri in Ubiri ward, and
Soni and Shashui villages in Soni ward in Lushoto district in the northeastern part of Tanzania.
The exiting land use pattern in the district is divided into four categories namely dry land farming
constituting 58%, tree crops or irrigated area (11%), forest reserves (16%) and grazing areas
(15%). This land use pattern makes the West Usambaras one of the most intensely farmed areas
of Tanzania (Pfeiffer, 1990).

A cross-sectional research design was adopted in this study because it allows data to be collected
at a single point in time without repetition from the representative sample. The reason for the
choice of such a design is that, it is easier and economical to conduct especially where resource
constraints like time, labour and money dictate the results, as it was the case for this study.
Primary data were collected through household questionnaire survey using structured
questionnaires with both open-ended and closed-ended questions, focus group discussion with the
key informants using a checklist of questions and physical observation, while secondary data
were collected through documentary review. The aim was to cross check and verify information
obtained through these different methods regarding the topic in question.

The target group for this study was all farmers practicing and not practicing agroforestry in the
Lushoto district. The sampling unit was the household. A household in this study is referred to as
a single person or group of persons who live and eat together and share common living
arrangements i.e. share expenses (URT, 1994). Since for any land use system to qualify as an
agroforestry system, there must be woody perennials (trees, shrubs etc) deliberately combined
with herbaceous plants and or livestock in spatial, rotational or both in which there are both
ecological and economical interactions between the tree and non-tree components of the system
(Young, 1989; ICRAF, 1996). In addition, since it was also hardly possible to find a single
household’s farm without trees and /or shrubs in the study area, then, the criterion used to obtain
farmers practicing and not practicing agroforestry was based on the number of trees found in the
household’s farm. Farmers who had less than 20 trees/ ha were regarded as not practicing
agroforestry while those with more than 20 trees were practicing agroforestry.
A Monthly Double-Blind Peer Reviewed Refereed Open Access International e-Journal - Included in the International Serial Directories
Indexed & Listed at: Ulrich's Periodicals Directory ©, U.S.A., Open J-Gage, India as well as in Cabell’s Directories of Publishing Opportunities, U.S.A.
International Journal of Physical and Social Sciences
http://www.ijmra.us
372
July IJPSS Volume 2, Issue 7 ISSN: 2249-5894

2012

_________________________________________________________

It was assumed that, other factors being constant, farmers with many agroforestry trees and which
are properly arranged and managed could have higher income resulted from positive interactions
between the tree and non-tree components of the system(s) than those with fewer trees.

After identifying farmers practicing and not practicing agroforestry; simple random sampling
technique was applied to select households from the study villages. Equal number of households
from both farmers practicing and not practicing agroforestry in each village were picked for the
interview for easy comparison. A random sampling intensity of at least 5% was used to determine
the sample size of the households interviewed in each village as described by Boyd et al. (1981)
cited by Kayunze (1998). Overall, a total of 134 households were selected for the study.

This study was carried out in two phases. Phase one involved a reconnaissance survey. The
second phase was mainly based on questionnaire survey. The reconnaissance survey was
conducted in order to observe the general conditions of the farming systems; make researcher be
acquainted with the study area; select study villages and the sample size required; and pre-test the
questionnaires to check for its validity and reliability, to fit the local condition as recommended
by Goldman and Macdonald (1987) cited by Kayunze (1998). Questionnaires were pre-tested
using 44 respondents from four villages. Eight respondents (4 from participants and 4 non-
participants of the agroforestry practices) and 3 key informants were picked from each village.
However, most of the questions were responded thus, very little modifications were made to the
original questions.

Descriptive and inferential statistics methods were applied to analyse the quantitative data while
qualitative data were analysed by using Content and Structural-Functional methods (Merina,
2001). Multiple linear regression model was developed in order to predict whether or not the
dependent and independent variables were significantly related and measure the strength of their
relationship. The dependent variable, net income of household’s farm production was regressed
on the independent variables (farm size, household size, number of trees planted, number of
livestock kept, varieties of crops grown and cost of production) to find the standard regression
A Monthly Double-Blind Peer Reviewed Refereed Open Access International e-Journal - Included in the International Serial Directories
Indexed & Listed at: Ulrich's Periodicals Directory ©, U.S.A., Open J-Gage, India as well as in Cabell’s Directories of Publishing Opportunities, U.S.A.
International Journal of Physical and Social Sciences
http://www.ijmra.us
373
July IJPSS Volume 2, Issue 7 ISSN: 2249-5894

2012

_________________________________________________________

coefficient, the beta weight (ß) of each independent variables, the multiple correlation, R, and the
multiple coefficient of determination, R2. These six independent variables were included because
they were thought to be able to account for more of the variation in the dependent variable. The
general model used in linear regression was:

Y = a + b x + b x + ……………+ bx + e
i 1 1 2 2 j j i

Where:
Y = The ith observed value of the net income of the household’s farm
i
production (dependent variable).
a = Intercept
b to b = Independent variable coefficients
1 6
X1 = Farm size
X2 = Household size
X3 = Number of trees planted
X4 = Number of animals kept
X5 = Varieties of crops grown
X6 = Cost of farm production
e = Random error
i

Results and Discussion
Socio-economic characteristics of farmers
The socio-economic characteristics of farmers examined in this study were sex, marital status,
age, household size, farm size, education level and the main occupation. The purpose of choosing
A Monthly Double-Blind Peer Reviewed Refereed Open Access International e-Journal - Included in the International Serial Directories
Indexed & Listed at: Ulrich's Periodicals Directory ©, U.S.A., Open J-Gage, India as well as in Cabell’s Directories of Publishing Opportunities, U.S.A.
International Journal of Physical and Social Sciences
http://www.ijmra.us
374
July IJPSS Volume 2, Issue 7 ISSN: 2249-5894

2012

_________________________________________________________

these characteristics was to get general overview of what the respondents are composed of and
how these characteristics could influence agroforestry practices towards poverty reduction in the
study area. Results in Table 1 show that, the highest percent of farmers (79.1%) were men while
only 20.9% were women despite the fact that women are the key players in most of the
household’s farm activities. The possible explanation for this trend is that the study targeted
households’ heads as main decision makers of the household affairs. Except for the few
households which were female headed, the majority were male headed. Sometimes, women had
to respond on behalf of their husbands due to some special excuses like when were not around.
Therefore, combining some of the households which were male headed and the female headed;
87.3% of the interviewed farmers were heads of the households while 12.7% not heads of the
households. The findings also show that 85.8% of the respondents were married, 6.7% divorced,
3.7% widowed and 3.7% separated.

The study found the age of farmers to range from 22–70 years. The average age was 45 years.
Over three quarters (79.1%) were in the age group of 31–64 years, whereas 12.7% were above 64
years and 8.2% below 30 years. This implies that most of farmers were in the economically
productive age group with great experience in agroforestry both before and after SECAP.
Therefore their experience was very useful in the success of this study. According to Mandara
(1998) and Mtenga (1999), household members are considered economically productive from the
age of 16 to 64 years. The age bracket below 16 years is children some of whom may be attending
schools and others too young to participate in farming activities. The age group above 64 years is
considered less economically active because the members are too old.

Table 1: Characteristics of respondents (n = 134)
Variable Characteristics Frequency Percent
Sex Male 106 79.1
Female 28 20.9

A Monthly Double-Blind Peer Reviewed Refereed Open Access International e-Journal - Included in the International Serial Directories
Indexed & Listed at: Ulrich's Periodicals Directory ©, U.S.A., Open J-Gage, India as well as in Cabell’s Directories of Publishing Opportunities, U.S.A.
International Journal of Physical and Social Sciences
http://www.ijmra.us
375
July IJPSS Volume 2, Issue 7 ISSN: 2249-5894

2012

_________________________________________________________

Marital status Married 115 85.8
Divorced 9 6.7
Widowed 5 3.7
Separated 5 3.7

Age group Below 30 years 11 8.2
Between 30 – 64 years 106 79.1
Above 64 years 17 12.7

Education level No formal education 5 3.7
Adult education 11 8.2
Primary education 96 71.6
Secondary education 19 14.2
Diploma 3 2.2

Farm size Less than 2.0 ha 51 38.0
2.0 – 3.0 ha 45 33.6
More than 3.0 ha 38 28.4

The mean household size was 6.6 persons. The smallest household had 3 persons while the largest
had 11. This figure is higher than the district average of 4.7 persons reported in the National
Population and Housing Census (URT, 2002) and lower than that of 8.8 persons reported by
Moshi (1997) in the west Usambaras. The average household size below the age of 16 years was
3.0 persons, 2.8 persons between 16 – 64 years and 0.8 persons above 64 years. This implies that
at least every household in the study area has an average of 2.8 persons who can actively
participate in farming activities.
A Monthly Double-Blind Peer Reviewed Refereed Open Access International e-Journal - Included in the International Serial Directories
Indexed & Listed at: Ulrich's Periodicals Directory ©, U.S.A., Open J-Gage, India as well as in Cabell’s Directories of Publishing Opportunities, U.S.A.
International Journal of Physical and Social Sciences
http://www.ijmra.us
376
July IJPSS Volume 2, Issue 7 ISSN: 2249-5894

2012

_________________________________________________________

Close to three quarters of respondents (71.6%) had primary education, 14.2% secondary, 8.2%
adult education, 3.7% no formal education while 2.2% diploma education. Generally, except for
the minority (3.7%) who had no formal education, most of the respondents were educated. This
implies that, introduction of various agroforestry innovations in the study area are likely to be
successfully adopted because the majority could not only be trained by the extensionists but also
read from books and newsletters and other sources of information.

With regard to farm size, over one third of respondents (38%) had farm sizes below or equal to 2
hectares. The mean farm size was found to be 3.1ha. The minimum and maximum farm sizes
were 0.7 and 4.5 hectares respectively. However, majority of the households with farm sizes
greater than 3 hectares had plots of woodlots some planted with black wattle (Acacia meansii) and
others planted various tree species for timber, firewood, poles and other building materials among
which being eucalyptus and grevillea. Further, observations have shown that there is high farm
fragmentation ought to be influenced by the former system of shifting cultivation. Farmers have
several farm plots of various shapes and sizes located into different places within and sometimes
even out of the sample village. Close to three thirds of respondents (61.2%) complained that their
farms were not enough. However, further observations have shown that the average farm size of
3.1 ha could be large enough to meet requirements of most of the reported household size of 6.6
persons. However, most farms were not well managed including adding enough manure and
planting improved crop seeds. Historically, the study area had faced serious soil erosion resulted
from poor soil conservation measures before SECAP thus leading to low soil fertility. Therefore,
if farmers could manage them properly as recommended, will improve the systems productivity
and thus meet most of the household daily requirements and surplus for cash income.

Productivity of the agroforestry systems
Various levels of farm production were observed in the study area between farmers practicing and
not practicing agroforestry as described in the subsections here under.

A Monthly Double-Blind Peer Reviewed Refereed Open Access International e-Journal - Included in the International Serial Directories
Indexed & Listed at: Ulrich's Periodicals Directory ©, U.S.A., Open J-Gage, India as well as in Cabell’s Directories of Publishing Opportunities, U.S.A.
International Journal of Physical and Social Sciences
http://www.ijmra.us
377
July IJPSS Volume 2, Issue 7 ISSN: 2249-5894

2012

_________________________________________________________

Animal husbandry
Findings in Tables 2 and 3 show the distribution of various livestock per household for farmers
practicing and those not practicing agroforestry respectively. Nearly every household (94%)
keeps at least one kind of livestock. Chicken was the most preferred livestock (88.8%), followed
by cows (46.5%), goats (43.8%), sheep (22.4%) and ducks (9.7%). Chicken was the most
preferred because it is relatively easy and cheap to manage as needs only small initial capital
compared to other livestock species. On average, every household had 9.7 chicken, 2.4 cows, 3.1
goats, 2.6 sheep and 1.2 ducks. The minimum and maximum number was 2 and 25 for chicken, 1
and 5 cows, 1 and 10 goats, I and 6 sheep, and 1 and 4 for ducks respectively.

Table 2: Average number of livestock kept per household for farmers practicing agroforestry
Village Average number per household
Cows Goats Sheep Chicken Ducks
Ubiri 2.7 3.3 3.3 10.5 2.0
Ngulwi 2.8 3.1 4.8 11.6 1.0
Soni 2.6 3.2 2.6 9.9 0.0
Shashui 2.8 3.9 4.0 9.1 2.0
Average 2.7 3.4 3.7 10.3 1.3
Generally, farmers practicing agroforestry had significantly higher numbers of livestock than
those not practicing (P<0.05). Sheep and ducks were the least preferred animals in the study area.
In Soni village for example, none of the farmers practicing agroforestry kept ducks (Table 2),
whereas in Ngulwi and Shashui villages results show that none of them was kept either sheep or
ducks among farmers not practicing agroforestry (Table 3).

Table 3: Average number of livestock per household for farmers’ not practicing agroforestry
Village Average number per household
A Monthly Double-Blind Peer Reviewed Refereed Open Access International e-Journal - Included in the International Serial Directories
Indexed & Listed at: Ulrich's Periodicals Directory ©, U.S.A., Open J-Gage, India as well as in Cabell’s Directories of Publishing Opportunities, U.S.A.
International Journal of Physical and Social Sciences
http://www.ijmra.us
378